FastLabel Inc. (Headquarters: Shinjuku-ku, Tokyo; President and CEO: Takeshi Suzuki; hereinafter "FastLabel") and ｕｇｏ Inc. (Headquarters: Chiyoda-ku, Tokyo; CEO: Ken Matsui; hereinafter "ｕｇｏ"), a developer of Japanese-made AI robots, announce the launch of the "ugo VLA Model Development Training Program powered by FastLabel," a hands-on training program utilizing the Japanese-made humanoid robot "ｕｇｏ Pro R&D Model" to enable companies, universities, research institutions, and other organizations to engage in Vision-Language-Action (VLA) model development from the early stages. Through this training program, participants can experience the entire process, from task design, imitation learning data collection, data quality management, and VLA model development (fine-tuning) to evaluation using an actual robot and reporting. Even organizations that have not yet established in-house knowledge or a development framework for Physical AI can experience the development and validation process using a robot arm in a short period of time, enabling them to conduct concrete evaluations for the adoption of Physical AI within their own organizations. BackgroundIn recent years, against the backdrop of Japan's aging population and labor shortages, improving operational efficiency and reducing workforce requirements in on-site operations such as security, inspection, transportation, and customer guidance have become important social challenges. In this context, Physical AI, including Vision-Language-Action (VLA) models that integrate vision, language, and action to enable flexible robot control, has attracted increasing attention as a next-generation technology capable of adapting to diverse real-world environments that are difficult to address through conventional rule-based automation.At the same time, organizations seeking to explore the use of Physical AI require an environment in which they can seamlessly evaluate the entire development process—from procuring robot hardware and collecting imitation learning data to model development and evaluation on actual robots. In practice, however, many organizations face challenges such as not knowing where to begin, lacking in-house expertise for development and validation, and facing a high initial burden before launching a proof of concept (PoC). These challenges have become major barriers to adoption in the early stages.To address these challenges, ｕｇｏ and FastLabel believe it is important to provide an environment that enables organizations to take the "first step" toward adopting Physical AI. The two companies have jointly planned and developed a corporate training program that integrates Japanese-made humanoid robot hardware, data collection and preparation, VLA model development, evaluation, lectures, and facility tours. Through this training program, participants can go beyond theoretical consideration and gain a concrete understanding of the potential of Physical AI and the key considerations for applying it within their own organizations by experiencing the development and validation process using actual robots. Program OverviewThe "ugo VLA Model Development Training Program" comprehensively packages the processes required for the initial validation of VLA model development. It provides an end-to-end experience covering the setup of the ｕｇｏ Pro R&D Model, development environment configuration, task design, imitation learning data collection, data quality management, VLA model training, deployment, evaluation using an actual robot, analysis, and reporting. Through lectures as well as tours and hands-on demonstrations using the actual robot, the program is designed to enable organizations to build both foundational knowledge and practical expertise in Physical AI. This training program places a strong emphasis on hands-on support with a view toward future in-house development and decisions regarding full-scale deployment. By making full use of the best practices and development frameworks for VLA model development accumulated by ｕｇｏ and FastLabel, organizations can aim to complete the initial development and validation process and build internal knowledge in as little as approximately three months from the initial inquiry. Outlookｕｇｏ and FastLabel have been collaborating through research and development as well as business development activities in the field of AI robotics. Through the launch of this training program, the two companies will create an environment that enables more organizations to begin the initial validation of Physical AI development and contribute to expanding the potential for robot adoption in Japan. CommentsKen MatsuiPresident and CEO, ｕｇｏ Inc."Accelerating the social implementation of Physical AI requires more than simply developing high-performance models. We believe it is equally essential to provide an environment where the cycle of data collection, evaluation, and improvement using actual robots can be repeatedly experienced in conditions that closely resemble real-world operational environments. However, in Japan, there are still not enough environments where VLA model development can be validated using actual robot hardware, nor are there sufficient numbers of professionals with practical expertise in this field.Through this program, by combining the Japanese-made humanoid robot ｕｇｏ Pro R&D Model with FastLabel's data platform and development expertise, we aim to provide companies and research institutions with an environment where they can practically learn and validate Physical AI development. We hope to contribute to cultivating talent capable of developing Physical AI for real-world applications and to building implementation know-how for Physical AI in Japan."Takeshi SuzukiPresident and CEO, FastLabel Inc."At physical AI development sites, there is a major shortage of human resources in Japan capable of handling new technology domains, including VLA models. As this domain is implemented into society and takes off in the industry moving forward, we believe that the thin layer of practical human resources who can handle everything from data collection design to actual machine evaluation and on-site integration will become a bottleneck. While we typically provide hands-on support for each company's physical AI utilization through building data infrastructure and supporting model development, we simultaneously view it as equally important that knowledge and human resources are accumulated within each company. Through this program, by combining ｕｇｏ's domestic actual machines with FastLabel's data infrastructure, we will arrange an environment where each company can verify physical AI while directly facing their own operational sites and tasks, thereby contributing to raising the baseline of the entire Japanese AI robotics industry." About FastLabelFastLabel, Inc. is committed to building data infrastructure that enables data-centric AI development. The company provides end-to-end support across the AI development lifecycle, including data collection and generation, annotation, model development, and DataOps implementation.In recent years, FastLabel has expanded its focus to the field of physical AI, including robotics, and has been developing data pipelines that support the creation of robot foundation models and Vision-Language-Action (VLA) models. Company OverviewCompany Name: FastLabel Inc.Headquarters: Shinjuku Sumitomo Building 24F, 2-6-1 Nishi-Shinjuku, Shinjuku-ku, Tokyo, JapanRepresentative: Takeshi Suzuki, President and CEOBusiness: Provider of professional services and products supporting Data-centric AI developmentWebsite: https://fastlabel.ai/service/robotics Media ContactCorporate Communications FastLabel Inc. E-mail: pr@fastlabel.ai